Though the causes of ischemic stroke are often identified, the ~35% of patients without a known cause are labeled as cryptogenic strokes. Since there are no current methods for identifying causes of cryptogenic strokes, our laboratory has used a molecular approach in an attempt to solve this problem. We have examined gene expression in whole blood using whole genome microarrays and shown specific gene expression profiles in whole blood of patients who have known causes of stroke due to cardioembolism due to paroxysmal atrial fibrillation (PAF), non-PAF cardioembolism and large vessel atherosclerosis. PAF often causes a blood clot in the heart that embolizes to brain to produce a stroke. However, by the time an EKG and/or Holter monitor are performed after the stroke, the heart rhythm has returned to normal with no evidence of PAF and the stroke is designated as cryptogenic. In the following aims we test whether our gene expression profiles for known PAF causes of cardioembolic stroke will detect cryptogenic strokes caused by PAF cardioembolism. In addition, we will examine gene expression differences in women and men following strokes caused by PAF because women are at a significantly higher risk for PAF related stroke than men, and the biological reasons for this are not understood. Aim #1. Predict which cryptogenic strokes are caused by PAF using qRT-PCR measurement of RNA levels of genes identified in our previous microarray studies of strokes caused by PAF. Aim #2. Demonstrate the cryptogenic cortical strokes predicted to be caused by PAF in Aim #1 have in fact: (a) PAF on prolonged outpatient cardiac monitoring after the stroke; (b) or have PAF on repeated Holter monitoring after the stroke(c) or PAF on repeat EKGs after the stroke. Aim #3. Determine the gene expression differences in women compared to men following cryptogenic strokes caused by PAF using the subjects proven to have PAF in Aim#2. significance. The first two aims will predict PAF causes of cryptogenic strokes and confirm these by cardiac monitoring. The third aim will address molecular pathways that are different between women and men with PAF to begin identifying the biological factors associated with the increased risk of PAF related stroke in women. There are over 260,000 cryptogenic strokes/ year in the US which compares to 270,000 new breast cancer cases/ year and 50,000 new Parkinson's cases/year. Our data suggest that >50% of cryptogenic cortical strokes may be due to unrecognized cardioembolism and half of these would be due to PAF. If we could identify all of the PAF cardioembolic cryptogenic strokes, then these patients would be treated with coumadin, Dabigatran or other oral anti-coagulant instead of anti-platelet agents or nothing. This would prevent thousands of strokes per year and is equivalent to the number of strokes treated with tPA per year.